[CT] Use zips of SKPs to speed up "isolate tests" in recipes

scripts/slave/recipes/skia/ct_skps.py

Index: scripts/slave/recipes/skia/ct_skps.py
diff --git a/scripts/slave/recipes/skia/ct_skps.py b/scripts/slave/recipes/skia/ct_skps.py
index abe50e9991cb075e20eaf7fdf5ae2a4bf6f7f1c6..6f1f7a98cc6f91295db6519e43e90c4e3ea649d5 100644
--- a/scripts/slave/recipes/skia/ct_skps.py
+++ b/scripts/slave/recipes/skia/ct_skps.py
@@ -17,6 +17,7 @@ DEPS = [
   'swarming',
   'swarming_client',
   'depot_tools/tryserver',
+  'zip',
 ]
 
 
@@ -101,6 +102,7 @@ def RunSteps(api):
   skps_chromium_build = api.properties.get(
       'skps_chromium_build', DEFAULT_SKPS_CHROMIUM_BUILD)
   ct_num_slaves = api.properties.get('ct_num_slaves', DEFAULT_CT_NUM_SLAVES)
+  skps_dir = api.path['slave_build'].join('skps')
 
   # Set build property to make finding SKPs convenient.
   api.step.active_result.presentation.properties['Location of SKPs'] = (
@@ -113,8 +115,12 @@ def RunSteps(api):
   for slave_num in range(1, ct_num_slaves + 1):
     # Download SKPs.
     api.ct_swarming.download_skps(
-        ct_page_type, slave_num, skps_chromium_build,
-        api.path['slave_build'].join('skps'))
+        ct_page_type, slave_num, skps_chromium_build, skps_dir)
+
+    # Zip up the SKPs due to https://github.com/luci/luci-go/issues/9
+    api.zip.directory(step_name='zip skps dir',
+                      directory=skps_dir.join('slave%s' % slave_num),
+                      output=skps_dir.join('slave%s.zip' % slave_num))

[borenet, 2016/02/11 20:11:55]

I'm confused; what's the point of zipping the SKPs if you aren't using the zip
file?

[rmistry, 2016/02/11 20:42:53]

It is used by the .isolate file in the CL referenced in the description.

 
     # Create this slave's isolated.gen.json file to use for batcharchiving.
     isolate_dir = chromium_checkout.join('chrome')